Development and use of the Lives Saved Tool: a model to estimate the impact of scaling up proven interventions on maternal, neonatal and child mortality

avoid the misunderstanding highlighted by Chopra et al.—LiST should not only be used to capture ‘gains’ (lives saved) but also possible ‘losses’ (lives lost) as well. Capturing losses illustrates the importance of maintaining a system (e.g. immunization) or continuing life-saving household/family behaviours (e.g. exclusive breastfeeding, as countries may experience widespread behaviour changes with urbanization). Expressing ‘lives lost’ will allow users to also better understand the importance of continuing to invest in mature, population-based preventive interventions with high coverage. To illustrate the ‘lives lost’ potential of the tool, we have run LiST first as is normally done (left) and then in the opposite direction (right), using South Africa as an example with coverage (based on South Africa’s reported coverage to illustrate the point in Chopra et al.) for the third dose of diphtheria–tetanus– pertussis (DTP) coverage (Figure 1a and b). Running LiST in the opposite direction gives users an idea of the lives already being saved by ‘successful’ interventions—and highlights the importance of continuing to invest in these interventions, to maintain as well as further increase their coverage. When running LiST in the opposite direction, users should understand that the tool uses a proportional cause-specific (not intervention-focused) mortality calculation. The calculation first takes into account preventive measures (and assigns a protective factor in the formula), and then considers curative measures. When coverage of preventive interventions is reduced (e.g. running LiST in the opposite direction for DTP3), greater weight is placed on the potential impact of curative interventions—which will then diminish the effect on estimates attributable to ‘lives lost’. Without thorough guidance about how to modify or use the tool, so that a typical user can model the number of lives being saved with continued successful use of an intervention at baseline (by showing ‘lives lost’) plus additional lives saved if coverage increases, we fear that LiST will be viewed by decision makers as ‘turn-key’. It could also run the risk of being used as a stand-alone tool rather than as part of a series of planning tools as apparently intended. It is important for users to look at the comprehensive effects of prioritization with the tool—for new and low-coverage interventions that need to be strengthened and high-coverage interventions that need to be maintained and further strengthened. Furthermore, LiST does not assess health system constraints, thus giving no idea of the cost implications of increased coverage, that is, how credible a target is given the weaknesses of the current health system. Similarly, LiST does not distinguish between achieving mortality reductions with relatively cost-effective preventive interventions such as immunization or vitamin A supplementation vs reliance on more costly curative interventions. We encourage the LiST authors to incorporate clearer instructions into the tool so that the typical user can operate LiST at its full power—to express lives saved as well as lives lost if gains in higher coverage interventions are not maintained.